A retrospective multicentre clinical study on management of isolated splenic vein thrombosis: risks and benefits of anticoagulation

Abstract

Abstract Introduction Isolated splenic vein thrombosis (iSVT) is a common complication of pancreatic disease. Whilst patients remain asymptomatic, there is a risk of sinistral portal hypertension and subsequent bleeding from gastric varices if recanalisation does not occur. There is wide variation of iSVT treatment, even within single centres. We report outcomes of iSVT from tertiary referral hepatobiliary and pancreatic (HPB) units including the impact of anticoagulation on recanalisation rates and subsequent variceal bleeding risk. Methods A retrospective cohort study including all patients diagnosed with iSVT on contrast-enhanced CT scan abdomen and pelvis between 2011 and 2019 from two institutions. Patients with both SVT and portal vein thrombosis at diagnosis and isolated splenic vein thrombosis secondary to malignancy were excluded. The outcomes of anticoagulation, recanalisation rates, risk of bleeding and progression to portal vein thrombosis were examined using CT scan abdomen and pelvis with contrast. Results Ninety-eight patients with iSVT were included, of which 39 patients received anticoagulation (40%). The most common cause of iSVT was acute pancreatitis n = 88 (90%). The recanalisation rate in the anticoagulation group was 46% vs 15% in patients receiving no anticoagulation (p = 0.0008, OR = 4.7, 95% CI 1.775 to 11.72). Upper abdominal vascular collaterals (demonstrated on CT scan angiography) were significantly less amongst patients who received anticoagulation treatment (p = 0.03, OR = 0.4, 95% CI 0.1736 to 0.9288). The overall rate of upper GI variceal-related bleeding was 3% (n = 3/98) and it was independent of anticoagulation treatment. Two of the patients received therapeutic anticoagulation. Conclusion The current data supports that therapeutic anticoagulation is associated with a statistically significant increase in recanalisation rates of the splenic vein, with a subsequent reduction in radiological left-sided portal hypertension. However, all patients had a very low risk of variceal bleeding regardless of anticoagulation. The findings from this retrospective study should merit further investigation in large-scale randomised clinical trials.